Move two: establishing a niche

IBÉRICA 15 [2008]: 25-50

AbstractThe significant purpose of the author in the Introduction of a research article isto convince the reader about the importance of the work to be presented. Toachieve this end, a convincing “niche” needs to be built by evaluating, rejectingor indicating gaps in previous related work. The purpose of “establishing aniche” is to emphasize the current research project presented by the author. Thepresent paper investigates how Computer scientists use this obligatory step of“Create a Research Space” (CARS) model (Swales & Feak, 1994 & 2004) tohighlight their own research work. This paper not only compares the results withother similar studies but also presents an in-depth analysis of various types ofgap statements used in Computer Science research article Introductions. Theissue of cyclicity of this step and the linguistic indicators used for theestablishment of “niche” (the gap statements) are both discussed.

Key words: corpus-based study, CARS, niche, Computer Science, gaps.

ResumenSub-secci�n dos: Òcreando el nichoÓ

El autor de un artículo de investigación persigue, como finalidad principal en laintroducción de dicho artículo, convencer al lector sobre la importancia de sutrabajo. Para alcanzar su objetivo, habrá de “crear un nicho”, con lo que evaluará,rechazará o señalará las carencias encontradas en trabajos anteriores yrelacionados con el que se presenta. La elaboración o construcción de un“nicho” tiene por objeto resaltar la investigación que en ese mismo momentoestá exponiendo el autor. En el presente artículo se investiga cómo losespecialistas del campo de la informática realizan esta tarea obligatoriaajustándose al modelo de “crear un espacio de investigación” (Modelo CARS)propuesto por Swales & Feak (1994 & 2004) y así hacer notorio su trabajo de

Move two: establishing a nicheWasima ShehzadUniversity College Yanbu (Kingdom of Saudi Arabia)[email protected]

25

03 SHEHZAD.qxp 14/3/08 17:33 Página 25

IBÉRICA 15 [2008]: 25-50

WASIMA SHEHZAD

investigación. En este artículo se comparan los resultados con otros estudiosanálogos y, además, se analizan de forma pormenorizada los distintos tipos deenunciados que denuncian carencias (gap statements) y que son característicos delas introducciones de los artículos de investigación en informática. Asimismo sedelibera sobre el asunto de la “ciclicidad” presente en este paso del modelo y losindicadores lingüísticos de carencias que con más frecuencia se utilizan para laconstrucción del “nicho”.

Palabras clave: estudios de corpus, modelo CARS, nicho, informática,carencias.

1. IntroductionThe writing of the introduction section of a research article has been studiedby Swales (1981, 1983, 1984 & 1990), Cooper (1985), Swales & Najjar (1987),Hughes (1989), Fredrickson & Swales (1994), Lindeberg (1994), Posteguillo(1995 & 1999), Anthony (1999 & 2002), Hyland (1999a, 1999b & 2000), Kuo(1999), Lewin et al. (2001), and Shehzad (2006 & 2007), among others. Outof these, the works of Cooper (1985), Posteguillo (1996 & 1999), Posteguilloet al. (1998), Anthony (1999 & 2002), and Shehzad (2006 & 2007) havefocused on the discipline of Computer Science (CS) research articles.However, Cooper (1985) and Anthony (1999) have used a relatively smallercorpus (12 research articles) and the focus of Posteguillo (1995, 1996 &1999); Posteguillo et al. (1998) and Anthony (2001 & 2002) has been on theoverall schematic structure of research articles, titles and abstracts, whereasShehzad’s (2006) focus is on Move One and Shehzad (2007) concentrates on“indicating the structure” part of the introduction. Furthermore, none ofthese scholars has investigated in detail Move Two: “Establishing a niche”(indicating a gap in the previous research or extending previous knowledgein some way) of the “Create A Research Space” (CARS) model when writingthe introduction of a research article/paper, which is often thought to beobligatory in major research articles (Swales & Feak, 1994 & 2004), in detail.

In the CARS model, Move 2 is the key move as it is the hinge that connectsMove 1 to Move 3 –i.e., “what has been done” to “what the present researchis about”. It functions as a “mini-critique” and often consists of not morethan a sentence. By indicating a knowledge gap, the writer builds up a“demand” for the current contribution. “Essentially the gap represents anunresolved question that the current contribution seeks to solve”(Lindeberg, 2004: 89). The present study aims to bring forward different

26


ways through which this move is realized in CS research articles for acomprehensive understanding of Computer scientists’ rhetorical style rather.

2. Literature reviewThe research article is essentially a “product” in its finished form which isthe outcome of a complex process (Swales, 2004), introduction holding afundamental position in this process.

An introduction of a research article is a “crafted rhetorical artefact” and a“manifestation of rhetorical maneuver” (Swales, 1990: 155). Thismaneuvering also involves building up a convincing “niche” (Swales, 1990:142) through “elaborated criticisms or denials of previous knowledge claims”(Lindeberg, 1994: 138) –i.e. statements of gaps as indicated through Move 2of CARS model (Swales & Feak, 1994 & 2004). Lindeberg (2004: 40) usesthe term “foil” (a thin layer of bright metal placed under a displayed gem orpiece of jewelry to lend it brilliance) for gap statements which “emphaticallyforeground and enhance the current contribution by contrast”. Foil is also“A very thin layer of polished metal placed under a gem, especially an inferioror imitation gem, to give it more color or sparkle” (The World Book Dictionary,2001: 828) thus emphasizing its quality of setting off contrast through whicha “demand” for the current contribution is built up.

“Niche” is also a marketing term indicating the process of finding small butpotentially profitable market segments and designing custom-made productsto cater for these specialized markets. However, the “niche” should be largeenough to be profitable but ignored by the major industry players. Swales(1990) uses this term in its ecological sense describing the relational positionof a species or population in an ecosystem. The “niche” (Swales, 1990: 142)thus established includes how a population responds to the abundance of itsresources and enemies. Relating, both, the ecological and marketingconnotations to writing the introduction of a research paper, the “niche”provides a firm “background against which the magnitude or relevance ofthe current contribution stands out clearly” (Lindeberg, 2004: 89).

Another objective of displaying writers’ knowledge and understanding ofthe state of the art of their discipline is also achieved through gap statements(Bazerman, 1988). Unfortunately, the available literature about CS is limitedto the analysis of overall structure of the CS research article (Posteguillo,1995 & 1999) and the generic evaluation of the introduction of Software

MOVE TWO: ESTABLISHING A NICHE

IBÉRICA 15 [2008]: 25-50 27


Engineering article (Anthony, 1999). Hence, in the present paper I deal withthe pertinent question of how Computer scientists make use of gapstatements to build up a convincing argument in the “tightly contested”genre of research article.

3. MethodologyIn the previous works of this kind either researchers have limited their studies toa close analysis of a small number of texts such as Tarone et al.’s (1981 & 1998)detailed description of two astrophysics articles, or a general analysis of a largercorpus (Dudley-Evans’s (1984) corpus of 156 articles, or even a comparativeanalysis of the Hyland’s (2000) large corpus of 240 research articles from softand hard sciences). These approaches have their limitations (see Anthony, 2002).Therefore a balanced approach was adopted for the study in focus. A corpuscalled, Shehzad Computer Science Corpus (SCS Corpus) (Shehzad, 2005) of 56introductions of CS research articles from five different journals published bythe Institute of Electrical and Electronics Engineers (IEEE) Computer Societywas created (see Table 1 for detail). The initial intention though was to have 60texts but the December issues of four journals had not been published by thetime the corpus was created. The data was analysed with the help of computer-based tools such as WordSmith (Scott, 1997).

The remainder of the paper first presents a general description of Move 2in CS research articles introductions followed by taxonomy of the differenttypes of Move 2 identified in the SCS Corpus. The next section deals withthe linguistic indicators of gaps followed by a discussion about the cyclicalnature of this move.

4. Results and discussionThe first step of the rhetorical analysis was to identify the realization ofMoves which showed 94.64% occurrence of Move 2, thus Swales & Feak’s(2004) assertion of its being obligatory. This percentage is higher thanAnthony’s (1999) report (91.7%) which was much higher than Posteguillo’s(1995) results (57%) thus implying a growing trend in its usage. The reasoncould be the fact that because of increased research being conducted in thisdiscipline, Computer scientists feel a competitive pressure more than everand need to create a research space for their work.

WASIMA SHEHZAD

IBÉRICA 15 [2008]: 25-5028


More than the overall number of occurrences, the patterns of Move 2 werefound to be interesting as they were not only complex but also complicated.Before going into the details, a Journal-wise summary of Move 2 is given inTable 1.

Name of journal Abbreviation for journal Move 2occurrence Percentage

In comparison to a hundred percent presence of Move 2 in three journals81.81% occurrence in PAMI looks slightly lower, but still it is high enoughto generalize that Move 2 is an important part of the research papers in thisfield and Computer scientists use it quite frequently. In one extreme case, theintroduction opened with Move 2.

Example (1)

SOME recent investigations [1], [2] revealed that some of the majorproblems today in developing software systems are not technical butmanagerial. (INTR 25)

4.1 Taxonomy of Move 2

The writers show their knowledge of “the state of the art” of their field bystating a problem, a set of conflicting evidence or a question. On the basisof extensive elaborations of such gap statements made in the introductionsof CS, a taxonomy of Move 2 types has been prepared, defined anddiscussed as follows.

(a) Simple/Short

About 20% of the occurrences could be called “simple gap statements”because of their easy-to-follow and straightforward manner. As far as brevityis concerned, “establishing a niche” is supposed to be a brief transitoryperiod like the one between touching the water and starting kicking whileswimming. Swales & Feak (2004) thus recommend it to be fairly short. A fewinstances of brief Move 2s were found in the SCS Corpus.


IBÉRICA 15 [2008]: 25-50 29

IBÉRICA 15 [2008]: …-…4

4. Results and discussion

The first step of the rhetorical analysis was to identify the realization of Moves

which showed 94.64% occurrence of Move 2, thus Swales & Feak’s (2004)

assertion of its being obligatory. This percentage is higher than Anthony’s

(1999) report (91.7%) which was much higher than Posteguillo’s (1995) results

(57%) thus implying a growing trend in its usage. The reason could be the fact

that because of increased research being conducted in this discipline, Computer

scientists feel a competitive pressure more than ever and need to create a

research space for their work.

More than the overall number of occurrences, the patterns of Move 2 were found

to be interesting as they were not only complex but also complicated. Before

going into the details, a Journal-wise summary of Move 2 is given in Table 1.

Name of journal Abbreviationfor journal

Move 2occurrence

Percentage

Transactions on Computers ToC 10 90.90%

Transactions on Pattern and Machine Analysis PAMI 9 81.81%Transactions on Software Engineering SE 11 100%Transactions on Parallel and Distributed Systems PADS 11 100%Transactions on Knowledge and Data Engineering KDE 12 100%

Table 1. Journal-wise occurrences of Move 2 in the SCS Corpus.

In comparison to a hundred percent presence of Move 2 in three journals 81.81%

occurrence in PAMI looks slightly lower, but still it is high enough to generalize

that Move 2 is an important part of the research papers in this field and

Computer scientists use it quite frequently. In one extreme case, the introduction

opened with Move 2.

Example (1)

SOME recent investigations [1], [2] revealed that some of the major problems

today in developing software systems are not technical but managerial. (INTR 25)

4.1 Taxonomy of Move 2

The writers show their knowledge of “the state of the art” of their field by stating

a problem, a set of conflicting evidence or a question. On the basis of extensive

elaborations of such gap statements made in the introductions of CS, a taxonomy

of Move 2 types has been prepared, defined and discussed as follows.


Example (2)

Previous attempts to compare items from different ontologies have beenbased on an integrated ontology that was derived manually or semiautomatically. This work aims at … (INTR 51)

(b) Lengthy/Numbered

Contrary to Swales & Feak’s (2004) belief, that Move 2s are quite short, oftenconsisting of no more than a sentence, most of the Move 2s in CSintroductions are fairly lengthy. The few one sentence examples look morelike exceptions. An example of the Lengthy gap is given here.

Example (3)

These data sets are far too large to fit in main memory and are typicallystored in secondary storage devices. Linear scans are the only cost-effectiveaccess method; random access is prohibitively expensive. Some data sets,such as router packet statistics, meteorological data, and sensor network data,are transient and need not be realized on disk; the data must be processed asthey are produced and discarded in favor of summaries whenever possible.As the size of such data sets far exceeds the amount of space (main memory)available to an algorithm, it is not possible for a data stream algorithm to“remember” too much of the data scanned in the past. This scarcity of spacenecessitates the design of a novel kind of algorithm that …. Each scan of alarge set on a slow device is expensive, and so the criteria …. In the case of“transient” streams, i.e., … only one scan is possible. Furthermore, unlike

WASIMA SHEHZAD

IBÉRICA 15 [2008]: 25-5030


IBÉRICA 15 [2008]: …-… 5

Type Meaning Occurrence Percentage

Short/Simple brief, easy to identify 15 26%

Lengthy/Numbered extended to many sentences and paragraphs 14 25%

Embedded intertwined with Move One, step b. i.e.,“reviewing items of previous research”

13 23%

Multiple Act based on “acts” each realizing a sub-purpose 6 10%

Reported already identified by theresearchers/scholars of the target community

3 5%

ResearchQuestions/Rationale

purpose/research question/rationalestatement of the present research serving asthe gap statement

3 5%

Extension continuation/augmentation of the previousresearch

2 3%

Contrastive presentation of a comparableidea/algorithm/technique/method to theexisting works

2 3%

Table 2. Types of gap statements in CS RAs introductions.

(a) Simple/Short

About 20% of the occurrences could be called “simple gap statements” because

of their easy-to-follow and straightforward manner. As far as brevity is

concerned, “establishing a niche” is supposed to be a brief transitory period like

the one between touching the water and starting kicking while swimming.

Swales & Feak (2004) thus recommend it to be fairly short. A few instances of

brief Move 2s were found in the SCS Corpus.

Example (2)

Previous attempts to compare items from different ontologies have been based on

an integrated ontology that was derived manually or semi automatically. This

work aims at … (INTR 51)

(b) Lengthy/Numbered

Contrary to Swales & Feak’s (2004) belief, that Move 2s are quite short, often

consisting of no more than a sentence, most of the Move 2s in CS introductions

are fairly lengthy. The few one sentence examples look more like exceptions. An

example of the Lengthy gap is given here.

Example (3)

These data sets are far too large to fit in main memory and are typically stored in

secondary storage devices. Linear scans are the only cost-effective access method;

random access is prohibitively expensive. Some data sets, such as router packet

statistics, meteorological data, and sensor network data, are transient and need not

be realized on disk; the data must be processed as they are produced and discarded

in favor of summaries whenever possible. As the size of such data sets far exceeds

the amount of space (main memory) available to an algorithm, it is not possible


an online algorithm, a data stream algorithm may not be required to take anirrevocable action after the arrival of each point; it may be allowed to takeaction after a …... Clustering, a useful and ubiquitous tool in data analysis, is,in broad strokes, the problem of finding a partition of a data set so that,under some definition of “similarity,” similar items are in the same part ofthe partition and different items are in different parts. …(INTR 54)

Move 2 took about 45% of the space of this introduction. Anotherinteresting type of Move 2 is that one in which the authors use the discoursemarkers to logically explain the reasons why there is a need for the researchspace they intend to occupy.

Example (4)

Several characteristics of computers that execute such applications makethem unsuitable for storing sensitive information. First, the devices may bepoor and …. Second, can be easily stolen or …. Third, when data size becomelarge, storage management is expensive and prone to errors. (INTR 45).

Even in one case, the gap statement does not remain just a brief statementas suggested by Swales & Feak (1994 & 2004). It rather becomes a reportcovering over two pages and then mingling with Move 1, b. towards the end.Unlike casual reporting it begins in an organized manner and has beendivided into four points that are numbered.

Example (5)

However, most of the existing database models are designed under theassumptions that the data/information stored is precise and queries are crisp.In fact many of these assumptions are not valid for many of the nextgeneration information systems since they may involve complex informationwith uncertainty. In general, data/information in databases may be uncertainfor the following reasons:

A decision in ….Integrating data from …Information in some traditional …In natural languages, … (INTR 58)

(c) Embedded

Twenty two percent of Move 2s in CS research articles introductions wereembedded in Move 1, b. “reviewing items of previous research”. In mostinstances it is difficult to demark Move 2 from Move 1, b because of their


IBÉRICA 15 [2008]: 25-50 31


overlapping natures. Consider the example (6) in which a gap has beensupported by the relevant literature:

Example (6)

DNS-dispatcher-based systems can easily scale from locally to geographicallydistributed Web-server systems … [23]. However, dispatching requeststhrough the DNS has three problems that prevent … These issues have beenaddressed through ….[8], or multiple tiers of … [15]. This last approach hasthree main drawbacks. To avoid a system bottleneck at the A-DNS, the traffic… requires … Moreover, if any client request… [21]. Finally, the TTL period…. is considered too low. (INTR 40)

(d) Multiple Act

“Communicative purposes are realized in moves, which in turn, arecomposed of one or more act (each realizing a sub purpose)” (Lewin et al.2001: 44). Up to four to five acts in the realization of gap statements in theSCS Corpus have been noted. This shows that the purpose of “establishinga niche” may also have sub purposes depending upon the kind of researchone is involved with.

Example (7)

In high resolution palmprint images … many features that are similar tosingular points and minutiae points in a fingerprint can be obtained; however,these features cannot be observed in low-resolution images … Nevertheless,we can extract principal lines and wrinkles from …. In fact, line features playan important role in palmprint identification [3], [4]. Unfortunately, it isdifficult to obtain a high recognition rate using only principal lines becauseof … (INTR 21)

Example (8)

Despite the advantages of a NOW, the potential computation power of theseparallel architectures has not been exploited in depth. One reason is thedifficulty in developing …Traditionally a programmer must …. However,this approach is time consuming and impractical for …, particularly since…(INTR 38).

(e) Reported

In some cases, gaps are not identified by the authors, they are rather reportedas having been generally known or found by other researchers. So, instead of

WASIMA SHEHZAD

IBÉRICA 15 [2008]: 25-5032


pointing out the shortcomings, the writer “merely reports them second-hand, relying on previous authorities, presumably a less face-threatening act”(Lindeberg, 2004: 92). This is done in their effort of not to “tread on thetoes of other citizens of the academic universe” (Lewin et al., 2001: 43).Computer scientists also use this device sometimes.

Example (9)

Reference [10] shows that none of the existing architectures meets theuniversal access requirement. (INTR 29)

In such a statement, the author does not create a gap himself; he rather usesthe weaknesses and shortcomings found by the previous research, to serveas a background to present his own work.

Example (10)

Previous research [21], [23] suggests that cross-site communication andcoordination issues cause a substantial loss of development speed. (Move 2)In this paper, we investigate relationships among delay, communication,coordination and … (INTR 30)

Considering it a promotional strategy, Lindeberg (2004: 134) further statesthat “[t]he promotional effect of referring to outside sources as havingpointed out the gap is that the present writers’ claim of a gap is supportedby other researchers working in the same problem area and this presumablymakes the criticism less confrontational”.

(f) Research Question(s /Rationale

In this type of Move 2, a gap is established by posing research question(s)based on the weaknesses of the previous research.

Example (11)

Qualitative studies [21], [18] have shown how individuals are disrupted bycross-site coordination challenges. But question remains about thecumulative effects, for example, how distance affects the speed with whichsoftware engineering tasks are accomplished, and how distance is related toother important variables that influence speed, such as the size of a task, orthe number of people involved. In addition to being important researchquestions, these are critical pragmatic issues as … (INTR 30)


IBÉRICA 15 [2008]: 25-50 33


In the following example, we notice the rationale of the need of “interagentcommunication and synchronization in multiagent applications” whichserves as a step forward to work on a new research area rather than bridgingan already existing gap.

Example (12)

In this paper, we address the issue of interagent communication andsynchronization…. (Move 3 a.) Interagent communication andsynchronization in multiagent applications is needed due to two reasons.First, the application may be inherently distributed in nature…. Second, theperformance of the application can be significantly improved by employinginteragent communication and synchronization techniques…. (Move 2) Wepropose …. (INTR 39)

(g) Extension

The gap created may not always be for a new research work, it could be theextension of the author’s previous work.

Example (13)

… this paper extends the logic language presented in previous work of ours[5], [6], [7] with a second order logic syntax … (INTR 57)

(h) Contrastive

Here Move 2 is realized in the form of a contrast with other work(s). Thisis mostly done when an improved version of the already existing design ortechnique is presented. In one introduction, the second occurrence of Move2 was seen as a contrastive statement to the preceding announcement ofMove 3, step a. Lexical indicators such as “unlike” are used to make thiscontrast.

Example (14)

Unlike existing algorithms, when making decisions about reducing orordering a test suite, our algorithms consider the complexities of MC/DC.(INTR 27)

It may be noted here that whereas this taxonomy of gap statements offers amyriad of choices to Computer scientists for the realization of Move 2, itmay make this genre tradition complex and lead to comprehensiondifficulties for the audience.

WASIMA SHEHZAD

IBÉRICA 15 [2008]: 25-5034


4.2 Linguistic indicators of gaps

The following linguistic indicators used for the establishment of a gap in theprevious research (Swales & Feak 1994 & 2004, and personal discussionswith Prof. Swales) were studied in the present Corpus. For the purpose ofdiscussion these have been classified into three categories.

(a) Contrastive statements: however; while; but; although; nevertheless; asopposed to; rather than; with a few exceptions.

(b) Quantifiers and quasi-negatives: limited; few; little.

(c) Negatives: none of; not been; no [work/research/data/study].

(a) Contrastive Statements

Sometimes to avoid negative or quasi-negative comments, the use of acontrastive statement such as “however” “while” “rather than” “as opposedto” “although” “considerable”, “with few exceptions”, “tended to”, etc., canbe useful. These are used to soften the negative impact of gap statements.Thereby, concordance search was conducted and studied to determine theirrole in the establishment of gap statements and has been listed in thedescending order of gap statements occurred in the SCS Corpus.

Table 3 illustrates significantly high usage of “however” for the enactmentof gaps as compared to the usage of “while” and “but” for such executions.The percentage of the occurrence of “nevertheless” is the same as that of“while”; however, having one gap statement in the Corpus, it could easily beneglected.

Conjunctions Hits Gap Statements Percentages

(a.1) However

“However” with 36 occurrences seems to be the obvious indicator toestablish a gap, but this word alone does not guarantee a clear indication of


IBÉRICA 15 [2008]: 25-50 35

WASIMA SHEHZAD

Conjunctions Hits Gap Statements Percentages

However 58 36 62%

While 41 14 34%But 46 11 23.91%Although 21 11 52.38%Nevertheless 3 1 33.33%

As opposed to 4 2 50%Rather than 14 1 7.14%With a few exceptions 1 1 100%

TOTAL 188 77 40.95%

Table 3. Contrastive statements used for Move 2.

(a.1) However

“However” with 36 hits seems to be the obvious indicator to establish a gap, but

this word alone does not guarantee a clear indication of the gap. It is the phrase

following “however” that contains and conveys the meaning, for example:

Example (15)

However, software visualizations are often too simplistic and lack visual clues for

the viewer to correctly interpret them. (INTR 33)

Example (16)

However, none of this work contains a complete and a systematic investigation of

different probabilistic modeling techniques for query approximation. (INTR 59)

Example (17)


the gap. It is the phrase following “however” that contains and conveys themeaning, for example:

Example (15)

However, software visualizations are often too simplistic and lack visual cluesfor the viewer to correctly interpret them. (INTR 33)

Example (16)

However, none of this work contains a complete and a systematicinvestigation of different probabilistic modeling techniques for queryapproximation. (INTR 59)

Example (17)

However, dispatching requests through the DNS has three problems thatprevent load balancing among the Web nodes: (INTR 40)

Interestingly, four instances of “however” represented Step a. “stating thenature of research” and two were used in the context of “Principalfindings”.

Example (18)

We will show below, however, that a straightforward remedy is to base thesimilarity measure on a suitable path metric. (INTR 23) (stating the nature ofresearch)

Example (19)

However, our algorithm took more time to find its better answers. (INTR 54)(Principal finding)

In one exceptional case it was used to state the “limitation” of the workundertaken.

Example (20)

… of packets dropped for networks with soft linklevel flow control thathandle congestion by dropping packets, however, this issue is not explored inthis work. (INTR 44)

WASIMA SHEHZAD

IBÉRICA 15 [2008]: 25-5036


(a.2) While

The next common linguistic item that played a part in the establishment ofthe knowledge gap was, “while”. Out of the 41 entries, 14 were used toconvey gaps.

Example (21)

While most existing works on MANET take flooding as a straightforwardand direct solution, we show in [15] that a blind flooding may result inexcessive redundancy, contention, and collision. (INTR 5)

Apart from the gap statements, an interesting juxtaposition was seen in theseentries. Two each were used to write the “limitations” and “value” of thecurrent work.

Example (22)

While an in-depth discussion of all possible applications is not possible, wenext briefly discuss work that relates to the applications we use to illustrateour optimization approach. (INTR 23)

Example (23)

These studies show that our testsuite reduction techniques can be effectivein reducing test suites while providing acceptable performance. (INTR 27)(Value)

(a.3) But

“But” usually signals two types of contrast: (i) “denial-of-expectation” (whatis expected after a reading of the first conjunct turns out not to be true froma reading of the second); and (ii) the “marker of semantic contrast” (Celce-Murcia & Larsen-Freeman, 1999). Some examples from the eleven instancesof “but” used for making gap statements are given below:

Example (24)

Qualitative studies (e.g., [21], [18]) have shown how individuals are disruptedby cross-site coordination challenges. But question remain about thecumulative effects, for example, how distance affects the speed … and how… (INTR 30)

Limited use of “but” to augment the “value” of their own work was alsonoticed.


IBÉRICA 15 [2008]: 25-50 37


Example (25)

After a thorough and detailed study … we have discovered centralized anddistributed schemes that not only achieve good results, but also guaranteestable performance. (INTR 40)

(a.4) Although

With eleven examples, “although” seems to be a popular choice after“however” and “while” for the expression of knowledge gap when theComputer scientists want to hedge the weaknesses identified in the alreadyused algorithms, methods developed or design mechanisms, etc.

Example (26)

Although a large part of the behavior of grouping algorithms is determinedby the behavior of the grouping measure, in this paper, we will focus on theconsiderations for the grouping framework. (INTR 16)

The use of “although” has not been restricted to contrastive statements. Anexample each of Move One step a. and step b. and Move Three step c. andstep d. were also found.

Example (27)

Although mobile ad hoc networks have been mainly used in militaryapplications, they are being increasingly used for civilian applications such asvirtual class rooms, wireless local area networks, and law enforcement.(INTR 37) (Move 1 step a)

Example (28)

The definition of … was first characterized … by Henzinger et al. [38],although the works of Munro and Patterson [64] and of Flajolet and Martin[23] predate this definition. (INTR 54) (Move 1 step b)

Example (29)

… we have found out that, although in some cases a distributed redirectionalgorithm may also achieve slightly worse performance than some centralizedalternatives it has three major advantages that make its use preferable …(INTR 40) (Move 3 step c)

WASIMA SHEHZAD

IBÉRICA 15 [2008]: 25-5038


Example (30)

… our heuristic approach performs better in terms of all the aboveparameters, although the percentage improvement is relatively low. (INTR 1)(Move 3 step d)

(a.5) Nevertheless

Out of the three, one clear-cut case of the use of “nevertheless” in gapstatements was:

Example (31)

Nevertheless, limited work has been reported on palmprint identification andverification, despite the importance of palmprint features. (INTR 21)

(a.6) As opposed to

This contrastive statement included two examples of gap and one of“Value”.

Example (32)

E-textiles impose specific challenges as opposed to other applications in thegeneral area of networked systems. (INTR 8)

(a.7) Rather than

Surprisingly, the entries of “rather than” were over shadowed with Move 3,having four examples of Step a., one of Step b. and two of Step d. The onlyclear example of gap statement is given here.

Example (33)

On the negative side, requiring explicit identification of states requires muchmore consistency from stakeholders when constructing scenarios and forcesthem to reason about their system in terms of state “rather than” sequencesof actions. (INTR 26)

(a.8) With a few exceptions

Writers often hedge their full commitment by referring to deficiencies in theresearch model, theory or method which may compromise the accuracy oftheir results, thus allowing “the prudent researcher to present results havinganticipated criticism of the premises or methods by which they were


IBÉRICA 15 [2008]: 25-50 39


achieved” (Hyland, 1998: 143). So the writers use phrases like “with a fewexceptions”, rather than “nothing is known about” or “we have noknowledge about”. As Hyland (1988: 145) points out “Again, this is a way oflimiting the damage that might result from a bald proposition; the hedgefunctioning to contextualize the claims made by suggesting parameterswithin which they might be accepted as true”. An example from the Corpusis given here.

Example (34)

In fact, with a few exceptions, the creators of Web pages create links to otherpages, usually with an idea in mind that the linked pages are …. (INTR 55)

(b) Quantifiers and quasi-negatives

The “most common way to indicate a gap is to use a ‘quasi-negative’subject”. (Swales & Feak, 2004: 258). Quantifiers, such as the following,could be helpful in pointing out the shortcomings of the previous researchand thus preparing ground to accommodate the current work. “Few” and“little”, although not morphologically negative, are negative in meaning andin syntactic behavior. The concordance results were studied for gapstatements and are depicted in Table 4.

Quantifiers Hits Gap Statements Percentages

“Little” with a 66.66% occurrence seems to be at the top, but “limited”having lesser percentage comparatively (with 17 occurrences) appears to bea more popular choice for the exercise of gap statements.

(b.1) Limited

One way of establishing gap is to show that other investigators’ work hasbeen limited in certain ways or to certain areas and thus intended to beencompassed by the present research. The word “limited” registered 28 hits.Despite having some open cases such as: “How do we search for a queriedpalmprint in a given database and obtain a response within a limited time?”,

WASIMA SHEHZAD

IBÉRICA 15 [2008]: 25-5040

IBÉRICA 15 [2008]: …-… 13

E-textiles impose specific challenges as opposed to other applications in the

general area of networked systems. (INTR 8)

(a.7) Rather than

Surprisingly, the entries of “rather than” were over shadowed with Move 3,

having four examples of Step a., one of Step b. and two of Step d. The only clear

example of gap statement is given here.

Example (33)

On the negative side, requiring explicit identification of states requires much more

consistency from stakeholders when constructing scenarios and forces them to

reason about their system in terms of state “rather than” sequences of actions.

(INTR 26)

(a.8) With a few exceptions

Writers often hedge their full commitment by referring to deficiencies in the

research model, theory or method which may compromise the accuracy of their

results, thus allowing “the prudent researcher to present results having

anticipated criticism of the premises or methods by which they were achieved”

(Hyland, 1998: 143). So the writers use phrases like “with a few exceptions”,

rather than “nothing is known about” or “we have no knowledge about”. As

Hyland (1988: 145) points out “Again, this is a way of limiting the damage that

might result from a bald proposition; the hedge functioning to contextualize the

claims made by suggesting parameters within which they might be accepted as

true”. An example from the Corpus is given here.

Example (34)

In fact, with a few exceptions, the creators of Web pages create links to other

pages, usually with an idea in mind that the linked pages are …. (INTR 55)

(b) Quantifiers and quasi-negatives

The “most common way to indicate a gap is to use a ‘quasi-negative’ subject”.

(Swales & Feak, 2004: 258). Quantifiers, such as the following, could be helpful

in pointing out the shortcomings of the previous research and thus preparing

ground to accommodate the current work. “Few” and “little”, although not

morphologically negative, are negative in meaning and in syntactic behavior.

The concordance results were studied for gap statements and are depicted in

Table 4.

Quantifiers Hits Gap Statements Percentages

Limited 28 17 60.71%Few 16 4 25%

Little 9 6 66.66%TOTAL 44 27 61.36%

Table 4. Gap statements and quantifiers/quasi-negatives.


17 clear examples were found that exhibited the limitations of the earlierresearchers. These included limitations in the designs, methods adopted,efficiency and accuracy, and limited work/research etc.

Example (35)

The disadvantages of the repeated application of the BOM test… are testtime inefficiency and “limited” fault coverage for …. This paper presents animproved… (INTR 10)

(b.2) Few

From the concordance list having 16 entries for “few”, only four could becategorized as gaps.

Example (36)

There are only a “few” studies to incorporate methods for handlingimpreciseness in such an integrated environment. (INTR 58)

(b.3) Little

The next quantifier looked for was “little” as according to Swales & Feak(2004), the chances of using a full negative like “no studies” are thatsomebody can find an exception to such a strong statement. Six instancesout of the nine entries for “little” had interesting nouns (work, research,information, detection effect, empirical information/evidence) accompaniedwith them to report insufficient job done by the other discourse communitydwellers.

Example (37)

… there is so far very “little” empirical evidence that communication ofdefect information among inspectors yields significant gains …. (INTR 29)

(c) Negatives

The third category of the linguistic indicators studied in the Corpus wasnegatives. “The chief use of negatives is directed at a proposition already inthe discourse” and are used more to respond than to initiate, Celce-Murcia &Larsen-Freeman (1999: 196) explain. They further elaborate that one mightassume that negative would receive prominent stress whereas it is reduced tomitigate the disagreement, presumably in the interest of maintaining socialharmony. Negatives are generally known to be used for pinpointing


IBÉRICA 15 [2008]: 25-50 41


something which is either faulty or has not been paid sufficient attention to.So, one would assume the use of negatives as a natural choice to pave way forone’s own work. Contrary to this speculation, the tendency of the usage ofnegatives is lower here, when compared to the usage of conjunctions andquantifiers/quasi-negatives for gap statements (see Table 5).

Negatives Hits Gap statements Percentages

(c.1) None of

In the six entries of none of, one looked like an expression of the “value”of the present work. In the rest of the examples, efforts to establish gaps caneasily be noticed.

Example (38)

… “none of ” the above-mentioned approaches can guarantee to find theglobal minimum … (INTR 23)

Referents of “None of ” used for gaps were: “the above-mentionedapproaches”, “this work”, “the published papers”, “the approaches above”,“group communication services”, or “existing architectures”.

(c.2) Not been

The three entries of “not been” were clear cut cases of gap statements: “hasnot been exploited in depth”, “has not been pursued in the literature untilrecently”, “has not been studied in the past”. This negative was accompaniedwith verbs such as “exploited”, “pursued”, “studied”, etc., used as presentperfect tense in the passive voice.

(c.3) No work/research/data/study

Contrary to the previous two negatives, the only instance of “no work”could be taken as a “value” statement.

Example (39)

To the best of our knowledge, no work has been published on the … (INTR 4)

WASIMA SHEHZAD

IBÉRICA 15 [2008]: 25-5042


IBÉRICA 15 [2008]: …-… 15

Negatives are generally known to be used for pinpointing something which is

either faulty or has not been paid sufficient attention to. So, one would assume

the use of negatives as a natural choice to pave way for one’s own work.

Contrary to this speculation, the tendency of the usage of negatives is lower here,

when compared to the usage of conjunctions and quantifiers/quasi-negatives for

gap statements (see Table 5).

Negatives Hits Gap statements Percentages

None of 6 5 83%

Not been 3 3 100%No work/research… 1 - 0%

TOTAL 10 8 80%

Table 5. Gap statements and negatives.

(c.1) None of

In the six entries of none of, one looked like an expression of the “value” of the

present work. In the rest of the examples, efforts to establish gaps can easily be

noticed.

Example (38)

… “none of” the above-mentioned approaches can guarantee to find the global

minimum … (INTR 23)

Referents of “None of” used for gaps were “the above-mentioned approaches”,

“this work”, “the published papers”, “the approaches above”, “group

communication services”, or “existing architectures”.

(c.2) Not been

The three entries of “not been” were clear cut cases of gap statements: “has not

been exploited in depth”, “has not been pursued in the literature until recently”,

“has not been studied in the past”. This negative was accompanied with verbs

such as “exploited”, “pursued”, “studied”, etc., used as present perfect tense in

the passive voice.

(c.3) No work/research/data/study

Contrary to the previous two negatives, the only instance of “no work” could be

taken as a “value” statement.

Example (39)

To the best of our knowledge, no work has been published on the … (INTR 4)

4.3 Overall indicators of gaps

The overall situation of the linguistic indicators found in Move 2 (i.e., gap


4.3 Overall indicators of gaps

The overall situation of the linguistic indicators found in Move 2 (i.e., gapstatements used by Computer scientists) is depicted in Table 6.

Linguistic indicator Hits Occurrence as gap PercentageThe most dominant

Out of the total hits for negatives, 80% serve the purpose of establishinggaps in the introductions, which is followed by 61.36% presence ofQuantifiers. On the other hand, the number of occurrences correspondingto conjunctions in gap statements is the highest but lower than thepercentages for negatives and quantifiers. This is because of the usage ofConjunctions in other moves too as we saw earlier on.

(a) Negative Verbs and Adjectives

Swales & Feak (2004) suggest twelve verbs and seven adjectives for studentsto tell if they are strongly negative, neutral or slightly negative. Theconcordance search showed that Computer scientists do not seem to beinterested in this list except for the rare use of “concentrated on” .

Example (40)

hence, people have concentrated on designing on heuristic algorithms fordesigning … for … and analyzing their performance. (INTR 37)

According to Swales & Feak (2004), five verbs (i.e., “misinterpreted”, “failedto consider”, “disregarded”, “ignored” and “overlooked”) are strongnegatives and thus should be avoided. Out of these five, the use of one verb(such as “misinterpreted”) has been found only once in the CS researcharticles introductions. But interestingly, neither of the other verbs consideredless negative were used by the Computer scientists. Similarly, three stronglynegative adjectives (“misguided”, “unconvincing” and “unsatisfactory”) werenot found in the Corpus, neither the rest of the adjectives except one rareoccurrence of “incomplete” as shown in example (41).


IBÉRICA 15 [2008]: 25-50 43

WASIMA SHEHZAD

IBÉRICA 15 [2008]: …-…16

statements used by Computer scientists) is depicted in Table 6.

Linguistic indicator Hits Occurrenceas gap

Percentage The mostdominant

Conjunctions 148 62 41.89% However

Quantifier 44 27 61.36% LimitedContrastive statements 48 15 31.25% AlthoughNegatives 10 8 80% None of

Table 6. Linguistic indicators of Move 2.

Out of the total hits for negatives, 80% serve the purpose of establishing gaps in

the introductions, which is followed by 61.36% presence of Quantifiers. On the

other hand, the number of occurrences corresponding to conjunctions in gap

statements is the highest but lower than the percentages for negatives and

quantifiers. This is because of the usage of Conjunctions in other moves too as

we saw earlier on.

(a) Negative Verbs and Adjectives

Swales & Feak (2004) suggest twelve verbs and seven adjectives for students to

tell if they are strongly negative, neutral or slightly negative. The concordance

search showed that Computer scientists do not seem to be interested in this list

except for the rare use of “concentrated on” .

Example (40)

hence, people have concentrated on designing on heuristic algorithms for

designing … for … and analyzing their performance. (INTR 37)

According to Swales & Feak (2004), five verbs (i.e., “misinterpreted”, “failed to

consider”, “disregarded”, “ignored” and “overlooked”) are strong negatives and

thus should be avoided. Out of these five, the use of one verb (such as

“misinterpreted”) has been found only once in the CS research articles

introductions. But interestingly, neither of the other verbs considered less

negative were used by the Computer scientists. Similarly, three strongly negative

adjectives (“misguided”, “unconvincing” and “unsatisfactory”) were not found in

the Corpus, neither the rest of the adjectives except one rare occurrence of

“incomplete” as shown in example (41).

Example (41)

Parnas [4] assessed that most legacy systems suffer from several typical problems,

including … and outdated, incomplete, or missing documentation. (INTR 33)

Hence, a word of caution in this regard for the teachers of CS rhetoric while

using general guide books for academic writing widely available in the market.


Example (41)

Parnas [4] assessed that most legacy systems suffer from several typicalproblems, including … and outdated, incomplete, or missing documentation.(INTR 33)

Hence, a word of caution in this regard for the teachers of CS rhetoric whileusing general guide books for academic writing widely available in themarket.

(b) Cyclicity

A fairly typical feature of gap statements in longer introductions is that theytend to occur at several different places, “being reintroduced as anotheraspect of the current research question is elaborated” (Lindeberg 2004: 91).Swales (1990: 158) acknowledges that “Niche-establishment does notnecessarily occur only at the end of a literature review, but may followreviews of individual items, so that cycles of Move 1, step 3 and Move 2recur” (see also Shehzad (2006) on cyclicity for “establishing a researchterritory move”).

In Posteguillo’s (1995: 67) study, 75% introductions showed a cyclical patternfor Move 2 as he informed that “[i]t is normal to find this move repeated aseries of instances throughout the same introduction, usually alternatingwith steps in move 1”. Although the occurrence of Move 2s in the presentcorpus ranged from one to seven, on an average it remained 1.84. This givesus the percentage of 51.78% which is much lower than Posteguillo’s figure.The lower percentage could be attributed to the fact that many Move 2s thatwere embedded in Move 1, b. were excluded in this calculation. Thecomparative cyclicity pattern of Move 2 (journal-wise gap statements) ispresented here.

Although a straight line of one time occurrence of Move 2s in manyintroductions is obvious from this graph, yet, the graph clearly reflectsvariations in the number of recurrences of Move 2 in different journals anddifferent introductions. Seven recurrences in PAMI is the highest numberbut most Move 2 recurrences fall between one and three giving an averageof 1.84 per introduction.

WASIMA SHEHZAD

IBÉRICA 15 [2008]: 25-5044


Example (42)

Move 2: An inadequate optimization criterion will not solve theapplication problem, no matter how easy it is to compute theoptimum. Conversely, sophisticated criteria that …are useless inpractice as well. For this reason, optimization approaches areattractive …

Move 3.a: In this paper, we introduce a novel optimization technique..Move 2: Numerous problems in computer vision, including partitioning

and grouping, lead to combinatorial optimization problems ofthis type.

Move 3.a: In contrast to related work, no specific assumptions are madewith respect to …. As a consequence, our approach covers…

Move 2: The absence of any specific assumptions about … as well as …listed above motivated this investigation.

Move 1, b: … approaches to efficiently compute good minimizers have along history in literature…

Move 2: Apart from …, none of the above-mentioned approaches canguarantee to find … and, in general, this goal is elusive due tothe combinatorial complexity of these minimization problems.


IBÉRICA 15 [2008]: 25-50 45

IBÉRICA 15 [2008]: …-… 17

for Move 2 as he informed that “[i]t is normal to find this move repeated a series

of instances throughout the same introduction, usually alternating with steps in

move 1”. Although the occurrence of Move 2s in the present corpus ranged from

one to seven, on an average it remained 1.84. This gives us the percentage of

51.78% which is much lower than Posteguillo’s figure. The lower percentage

could be attributed to the fact that many Move 2s that were embedded in Move

1, b. were excluded in this calculation. The comparative cyclicity pattern of

Move 2 (journal-wise gap statements) is presented here.

Graph.1. Comparative recurrences (journal-wise) of Move 2.

Although a straight line of one time occurrence of Move 2s in many

introductions is obvious from this graph, yet, the graph clearly reflects variations

in the number of recurrences of Move 2 in different journals and different

0

1

2

3

4

5

6

7

8

1 2 3 4 5 6 7 8 9 10 11 12

CS Research Articles Introductions

Num

ber

ofoc

curr

ence

s

ToC PAMI SE PADS KDE

IBÉRICA 15 [2008]: …-… 17

cyclicity for “establishing a research territory move”).

In Posteguillo’s (1995: 67) study, 75% introductions showed a cyclical pattern

for Move 2 as he informed that “[i]t is normal to find this move repeated a series

of instances throughout the same introduction, usually alternating with steps in

move 1”. Although the occurrence of Move 2s in the present corpus ranged from

one to seven, on an average it remained 1.84. This gives us the percentage of

51.78% which is much lower than Posteguillo’s figure. The lower percentage

could be attributed to the fact that many Move 2s that were embedded in Move

1, b. were excluded in this calculation. The comparative cyclicity pattern of

Move 2 (journal-wise gap statements) is presented here.

Graph.1. Comparative recurrences (journal-wise) of Move 2.

Although a straight line of one time occurrence of Move 2s in many

introductions is obvious from this graph, yet, the graph clearly reflects variations

in the number of recurrences of Move 2 in different journals and different

0

1

2

3

4

5

6

7

8

1 2 3 4 5 6 7 8 9 10 11 12

CS Research Articles Introductions

Num

ber

ofoc

curr

ence

s

ToC PAMI SE PADS KDE


Move 3.b: Consequently, the important question concerning the … Move 2: To the best of our knowledge, none of the approaches above

(apart from simulated annealing) seems to be immune against …and, hence, does not meet these criteria.

Move 3.a: Our approach belongs to …Move 2: Second, under certain conditions, bounds can be derived with

respect to the quality of the suboptimal solution. At present,these bounds are not tight with respect to the much betterperformance measured in practice.

Move 2: …these authors were able to show that … cannot be worse than… to the unknown global optimum. Besides … this fact hasmotivated our work. (INTR 23)

Another determining factor in the cyclicity of Move 2 is its physicalplacement in the introduction. The total number of words of eachintroduction was divided into four parts. Table 7 shows these linguisticindicators divided into four quarters of the introduction, in terms of theirphysical placement which reflects that most of the Move 2s occurred in thefirst three quarters of the introductions.Indicator Occurrence 1st quarter 2nd quarter 3rdquarter 4th quarter

Although the highest percentage is found in the third quarter of theintroductions, a fairly large percentage of gap indicators is present in the firstand second quarters also –hence, strengthening the proposition of thecyclicity of Move 2 in CS research articles introductions. Since this is apreliminary study, rather than resolving the issue of cyclicity, it raisesquestions for further research such as: why does this cyclicity occur? Why are

WASIMA SHEHZAD

IBÉRICA 15 [2008]: 25-5046


Indicator Occurrence 1st quarter 2nd quarter 3rd quarter 4th quarter

However 36 9 9 17 1

Limited 17 6 5 6 -While 14 3 8 3 -Although 11 2 2 7 -But 11 4 2 4 1

Little 6 2 1 3 -None of 5 - 2 3 -Few 4 - 2 2 -Not been 3 1 1 - 1

As opposed to 2 - - 2 -Rather than 1 - 1 - -With a fewexceptions

1 1 - - -

Nevertheless 1 1 - - -No work - - - - -

Total 112 29 33 47 3Percentage 25.89% 29.46% 41.96% 2.69%

Table 7. Physical placement of Move 2 indicators.

Although the highest percentage is found in the third quarter of the introductions,

a fairly large percentage of gap indicators is present in the first and second

quarters also –hence, strengthening the proposition of the cyclicity of Move 2 in

CS research articles introductions. Since this is a preliminary study, rather than

resolving the issue of cyclicity, it raises questions for further research such as:

why does this cyclicity occur? Why are journals such as PAMI more prone

towards this characteristic? Does the length of the article and conceptual profile

of the paper affect the pattern of Move 2?

5. Conclusion

Gap statements function as a contrasting background to strengthen the

significance of the present contribution. They serve as a stage with a brief

announcement of the new drama to be presented, thus raising the expectations

mingled with the expected applause of the audience. An important decision to be

made is how much prominence can be or should be afforded for this step as we


journals such as PAMI more prone towards this characteristic? Does thelength of the article and conceptual profile of the paper affect the patternof Move 2?

5. ConclusionGap statements function as a contrasting background to strengthen thesignificance of the present contribution. They serve as a stage with a briefannouncement of the new drama to be presented, thus raising theexpectations mingled with the expected applause of the audience. Animportant decision to be made is how much prominence can be or should beafforded for this step as we saw that gap statements in CS research articles arenot instances of brief announcements only, neither are these fixed in theirphysical or rhetorical nature. They can be found at different places in thetexts. These can vary from simple to complex, from creating a rationale (thatserves as a gap) to the extension of the previous work, and from contrastivestatements to reported gap statements. Conjunctions, quantifiers, negativesand contrasts are typically used as the signal of the onset of this move.

With rapid developments in CS research, Computer scientists have to find astrong “niche” to increase the chances of their audience’s acceptability andthe target community’s recognition of their research work which is essentialfor scientific progress. The editors and reviewers’ increasing pressures forreliable, quality controlled scholarly scientific information with their highquality standards for assessing research for publication in academic journalsmay be the compelling factor for the authors to use innovative strategies tocreate a “gap”, a “niche” so that they can provide the cognitive bridgebetween the known and unknown, old and new, faulty and improved,obsolete and latest, ineffective and efficient, slow and rapid, delayed andsubsequent etc., successfully. However, individual disciplinary practices whenwriting a research article may vary according to their communicativepurposes as the choice of the rhetorical strategies and linguistic indicatorsused, and the notion of cyclicity in the CS research article introductionsreflect. This selection is deliberate or content/audience specific needs to beresearched further.

Acknowledgements

I am grateful to Professor John Swales for his kind guidance, to Santiago


IBÉRICA 15 [2008]: 25-50 47


Posteguillo for sharing scholarly thoughts, and also to the anonymousreviewers for their invaluable insights.

(Revised paper received July 2007)

References

WASIMA SHEHZAD

IBÉRICA 15 [2008]: 25-5048

Anthony, L. (1999). “Writing re-search article introductions insoftware engineering. How ac-curate is a standard model?”IEEE transactions on Profes-sional Communication 42: 38-46.

Anthony, L. (2001). “Character-istic features of research articletitles in computer science”. IEEETransactions on ProfessionalCommunication, 44: 187-194.

Anthony, L. (2002). A MachineLearning System for the Auto-matic Identification of Text Struc-ture and Application to ResearchArticle Abstracts in ComputerScience. PhD Dissertation.Birmingham: The University ofBirmingham.

Bazerman, C. (1988). ShapingWritten Knowledge: The Genreand Activity of the ExperimentalArticle in Science. Madison, WI:The University of WisconsinPress.

Celce-Murcia, M. & D. Larsen-Freeman (1999). The GrammarBook, 2nd ed. Boston: Heinleand Heinle.

Cooper, C. (1985). Aspects ofArticle Introductions in IEEEPublications. MSc unpublisheddissertation. Birmingham: TheUniversity of Aston in Birming-ham.

Dudley-Evans, T. (1984). “A pre-liminary investigation of the writ-ing of dissertation titles” in G.James (ed.), The ESP Class-room Methodology, Materials,Expectations. 40-46. Exeter:University of Exeter.

Fredrickson, K.M. & J. Swales

(1994). “Competition and dis-course community: introductionsfrom Nysvenska studier” in B.L.Gunnarsson, P. Linell & B. Nord-berg (eds.), Text and Talk in Pro-fessional Context, 9-22. Upp-sala: ASLA.

Hughes, G. (1989). Article Intro-ductions in Computer Journals.Unpublished MA dissertation.Birmingham: The University ofBirmingham.

Hyland, K. (1998). “Boosting,hedging and negotiation of aca-demic knowledge”. Text 18: 349-382.

Hyland, K. (1999a). Writing:Texts, Processes and Practices.London: Longman.

Hyland, K. (1999b). “Disciplinarydiscourses: writer stance in re-search articles” in C. Candlin &K. Hyland (eds.) Writing: Texts,Processes and Practices, 99-121. London: Longman.

Hyland, K. (2000). DisciplinaryDiscourses: Social Interactionsin Academic Writing. Harlow, Es-sex: Longman.

Kuo, C.H. (1999). “The use ofpersonal pronouns: role relation-ships in scientific journal arti-cles”. English for Specific Pur-poses 18: 121-138.

Lewin, A., J. Fine & L. Young(2001). Expository Discourse: AGenre-based Approach to SocialScience Research Texts. NewYork: Continuum.

Lindeberg, A.C. (1994). “An ex-ploratory study of knowledgeclaims in article introductions inthree disciplines: Finance, Man-agement, and Marketing” in M.

Majapuro & T. Nikko (eds.)Talous ja kieli II (Business andLanguage II), 647-655. Helsinki:Publications of the HelsinkiSchool of Economics and Busi-ness Administration.

Lindeberg, A.C. (2004). Promo-tion and Politeness: ConflictingScholarly Rhetoric in Three Dis-ciplines. Vasa: Abo AkademiUniversity Press.

Posteguillo, S. (1995). GenreAnalysis in English for ComputerScience. Unpublished PhD dis-sertation. Valencia: Universitatde València.

Posteguillo, S. (1996). “A genre-based approach to the teachingof reading and writing abstractsin Computer Science” in J.Piqué, J.V. Andreu-Besó & D.J.Viera (eds.), English in SpecificSettings, 47-58. Valencia: NauLlibres.

Posteguillo, S. (1999). “Theschematic structure of computerscience research articles”. Eng-lish for Specific Purposes 18:139-160.

Posteguillo, S., I. Fortanet, J.C.Palmer & J.F. Coll (1998). “Lin-guistic analysis of research arti-cle titles: disciplinary variations”in I. Vázquez Orta & I. GuillénGalve (eds.), Perspectivas prag-máticas en linguística aplicada,443-458. Zaragoza: Universidadde Zaragoza.

Scott, M. (1997). WordSmithManual. Oxford: Oxford Univer-sity Press.

Shehzad, W (2005). Corpus-based Genre Analysis: Comput-er Science Research Article In-


Dr. Wasima Shehzad currently teaches linguistics at the University CollegeYanbu, KSA and has taught technical communication/writing at NationalUniversity of Sciences and Technology, Pakistan. Her research interestsinclude corpus-based genre analysis, academic/technical writing and teachereducation. Her recent publications include Teaching Communication Skills(2006), Bookman Publishers.


IBÉRICA 15 [2008]: 25-50 49

troductions. PhD Dissertation,Islamabad, National Universityof Modern Languages, Pakistan.

Shehzad, W. (2006). “Computerscientists’ approach to ‘estab-lishing a research territory’”. Se-lected Papers from the FifteenthInternational Symposium onEnglish Teaching, 127-141.Taipei: Crane Publishing Co andETA-ROC.

Shehzad, W. (2007). “How toend an introduction in a Comput-er Science article” in E. Fitz-patrick (ed.), Corpus LinguisticsBeyond the Word: Corpus Re-search from Phrase to Dis-course, 243-255. New York:Rodopi.

Swales, J.M. (1981). Aspects ofArticle Introductions. Aston ESPResearch Report No. 1, Lan-guage Studies Unit. Birming-ham: The University of Aston inBirmingham.

Swales, J.M. (1983). “Develop-ing materials for writing scholar-

ly introductions” in R.R. Jordan(ed.), Case Studies in ELT, 180-200. London: Collins.

Swales, J.M. (1984). “Researchinto the structure of introductionsto journal articles and its appli-cation to the teaching of aca-demic writing” in R. Williams, J.Swales & J. Kirkman (eds.),Common Ground: Shared Inter-ests in ESP and CommunicationStudies, 77-86. Oxford: Perga-mon.

Swales, J.M. (1990). GenreAnalysis: English in Academicand Research Settings. Cam-bridge: Cambridge UniversityPress.

Swales, J.M. (2004). “Then andnow: a reconsideration of thefirst corpus of scientific English”.Ibérica 8: 1-14.

Swales, J.M. & C. Feak (1994).Academic Writing for GraduateStudents: A Course for Nonna-tive Speakers of English. AnnArbor: The University of Michi-

gan Press.

Swales, J.M. & C. Feak (2004).Academic Writing for GraduateStudents. Essential Tasks andSkills, 2nd ed. Ann Arbor: TheUniversity of Michigan Press.

Swales, J.M. & H. Najjar (1987).“The writing of research articleintroductions”. Written Commu-nication 4: 175-192.

Tarone, E. S. Dwyer, S. Gillette& V. Icke (1981). “The use of thepassive in two astrophysics jour-nal papers”. English for SpecificPurposes 1: 123-140.

Tarone, E. S. Dwyer, S. Gillette& V. Icke (1998). “On the use ofthe passive and active voice inastrophysics journal papers:with extensions to other lan-guages and other fields”. Eng-lish for Specific Purposes 17:113-132.

_____ (2001). The World BookDictionary. Chicago: The WorldBook.


Move two: establishing a niche

Documents

Transcript of Move two: establishing a niche